3222 Scheme I
rylating agents. The dimer (pyrophosphate) was excluded as the intermediate in question by its inability to phosphorylate 3 under these conditions.
0 -
qC0jH qzII> CO,H
H$O'-
0 0
I/
I1
cop*oco2- 0-
0
2
0
CH,CI.
P
Acknowledgments. W e are pleased to thank Professors J. M. Jordan and L. T. Scott of U C L A for some of the materials used in this study. Financial support was provided by the Fulbright Commission's Program of Cultural Cooperation between the U S A and Spain and by the donors of the Petroleum Research Fund, administered by the American Chemical Society. References and Notes (1)A. J. Kirby and S.G. Warren, "The Organic Chemistry of Phosphorous", Elsevier, Amsterdam, 1967,Chapter 10,and references cited therein. (2) J. Rebek and F. Gavina, J. Am. Chern. Soc.. 97, 1591 (1975). (3) G.DiSabato and W. P. Jencks, J. Am. Chem. Soc., 83, 5500 (1961). (4)Lithia!ed. 1 % cross-linked polystyrene (F. Camps, J. Castells. M. J. Ferrando, and J. Font, TetrahedronLett1713 (1971))was carboxylated with CO? at -78" in THF. Acid prepared in this manner showed 1 mequiv of
0 solid
@-
C H 2 0 2 C C H 2 N L@-CH,OjCCHjh"P*O3
titratable acid functions per gram of resin: L. T. Scott and C. S.Sims, private communication. (5) J. M. Stewart and J. D. Young, "Solid Phase Peptide Synthesis", W. H. Freeman, San Francisco, Calif., 1969. (6)J. Rebek, D. Brown, and S. Zimmerman. J. Am. Chem. Soc.. 97, 454 (1975).Detection of intermediates in these cases was possible even when the two resins were physically separated by wire screens or sintered glass barriers. (7)Visiting Fulbright Scholar on leave from the University of Valencia. Valencia, Spain.
II I
J. Rebek,* F. Caviiia' Contribution No. 3460, Department of Chemistry University of California, Los Angeles Los Angeles, California 90024 Received February 24, 1975
04
An acyl phosphate was selected as a suitable polymerbound metaphosphate precursor. This choice was based on the work of J e n ~ k s who , ~ showed that the decomposition of acyl phosphates in media of low water or high salt concentrations produces pyrophosphate (Le., trapping of metaphosphate by phosphate). A polymer-bound acyl phosphate linkage was therefore likely to generate metaphosphate cleanly in the aprotic medium of the polystyrene matrix. The precursor was easily prepared, although in low (1020%) overall yield, by the reactions of Scheme I. The polymer-bound benzoic acid4 (ir 1720, 1670 cm-') was converted in large part to the anhydride (ir 1785, 1725 cm-I) with carbodiimide, then to the radioactive acyl phosphate 2 (ir 1725, 1230 cm-I) with tetramethylammonium [32P] phosphate in aqueous dioxane. That the phosphate of 2 was covalently bound to rather than adsorbed on the resin was demonstrated by its failure to exchange with unlabeled phosphate in solution. The trapping agent used was the polymer-bound glycine 3 (Scheme 11), prepared by established procedures of Merrifield peptide s y n t h e s i ~ . ~ When the polymers 2 and 3 were suspended in dioxane a t 80°, phosphate transfer between the two polymers was detected generating 4 (ir 1380 cm-'). Radioactivity assays indicated that the half-life of 2 is approximately 27 hr under these conditions and 70% of the released phosphate appeared on 4 while the remaining activity appeared in solution as phosphate. Saponification of 4 gave glycine N-phosphate, 5, identical with an authentic sample, and isotope dilution established that 90% of the radioactivity of 4 appeared as 5. Since direct reactions between the two resin bead surfaces have been shown to be negligible in related cases,6 the presence of a free monophosphorylating agent in the solution between the two solid phases is established. These results a r e consistent with the postulation of monomeric metaphosphate as the intermediate, but its higher oligomers, formed by disproportionation reactions within 2, remain viable alternative possibilities as the actual phosphoJournal of the American Chemical Society 1 97:Il
Does the Photochemical Bicyclopropenyl Rearrangement Involve a Prismane Intermediate?' Sir:
-
In his pioneering work on both thermal and photochemical versions of the bicyclopropenyl benzene rearrangement, Breslow suggested mechanism 1 with a prismane as the key intermediate to account for the observed ortho-para scrambling of x,y-substituents in the process of aromatization.* While we have dealt with the thermal (and transition metal catalyzed) case in recent paper^^-^ we have now turned to an investigation of the photochemical rearrangement whose mechanism has remained unchallenged so far.6 In order to facilitate product analysis by NMR, we modi-
May 28, 1975
Y'
?
X
3223 Scheme I
Ph
Ph
I
I
Ph
Ph
5
6
-L 7
Ph
Ph
\
Ph
7
Ph 4
'H
"72 -
L\= Ph
ph*I;
Ph Ph
c;#ph
Ph 10
I
Ph
H
6
8
fied Breslow's original compound l 2 by introducing a rnethyl group a t one of the bridgehead carbon atoms.'
- phAT\ \
Ph'
Ph
9
and 8 were formed but significantly no 5! (Reaction conditions and product analysis as above.) That is: prismane 3" does indeed behave according to eq 1 (ortho-para scrambling involving 9 and 10, see Scheme I) but it cannot be an intermediate in the photolysis of 2. Obviously there must exist a hitherto unrecognized pathway for photochemical bicyclopropenyl benzene rearrangements. In the accompanying paperI6 we report on observations which may be relevant in this context.
-
1
2
Acknowledgment. Support of this work by the Deutsche O u r r e ~ e n t discovery ~-~ of the A g k a t a l y z e d bicycloproForschungsgemeinschaft is gratefully acknowledged. penyl + Dewar benzene rearrangement now provides the basis for a crucial test of mechanism 1 (see Scheme I); the References and Notes supposed prismane intermediate 3 in the photolysis of 2 can (1) Reactions of coupled three-membered rings, part VII. For part VI of this also be independently reached via the conventional [ r S 2 series see R. Weiss, and H. P. Kempcke. Tetrahedron Lett. 155 (1974). x S 2 ]photocyclization of 9.4 As the further fate of 3 should (2) R. Breslow, P. Gal, H. W. Chang, and L. J. Altman, J. Am. Chem. Soc., 87, 5139 (1965). be independent of its mode of generation, one expects the (3) R. Weiss and C. Schlierf. Angew. Chem., lnt. Ed. Engl., 10, 811 (1971). same product distribution in the photolysis of both 2 and 9. (4) R. Weiss and S.Andrae. Angew. Chem., Int. Ed. Engl., 12, 150 (1973). (5) R. Weiss and S.Andrae, Angew. Chem., Int. Ed. Engl., 12, 152 (1973). This, however, proved to be not even qualitatively so. (6) G. Maier. "Vaienzisomerisierungen", Verlag Chemie, Weinheim, 1972, Direct irradiation of 2 a t 254 nm in benzene (Graentzel p 157. reactor 400,600 W , substrate concentration mol/l., ir(7) This was achieved by a conventional route, which, starting from I , involved hydride abstraction, quenching of the resulting cyclopropenium radiation time 5 hr) cleanly yielded benzene derivatives 5 ion2 with CH30H-CH30- and subsequent Grignard reaction with and 6 in a 1:2 ratio. No 8 could be detected. Product comCH3MgJ. (8) We thank Professor R. Breslow for experimental details concerning his position was established by VPC separation* and identificaVPC analysis of mixtures of isomeric tetraphenylbenzenes.* tion of the individual components by comparison with au(9) 6'O and e4 are known compounds and can easily be distinguished by thentic sample^.^ W e thus observe ortho-meta scrambling their NMR spectra and melting points (6, CH3 (s) T 7.79, mp 186': 8 CH3 (s) T 8.04, mp 270'). The nonidentity of 5 with either of these two of bicyclopropenyl bridgehead substituents in the process of isomers also follows from NMR and mp (5, CH3 (s) T 7.94, mp 226') aromatization!' This result strongly contrasts with Breswhile all three compounds have very similar uv and ir spectra as well as virtually identical fragmentation patterns in MS, which establishes their low's observation of ortho-para scrambling in photolysis of close structural relationship. The obvious suspicion that 5 should be as1 i 3 and is clearly inconsistent with (eq 1 ) as there is no set signed the structure of the hitherto unknown meta isomer in the series is of prismane F= Dewar benzene interconversions capable of confirmed by our observation that 5 is also formed-in addition to 8-in Ad-catalyzed aromatization of 9. We have found that this Ag'-induced this scrambling pattern. However, a prismane intermediate, meta-para scrambling of Dewar benzene bridgehead substituents in the 3, can still not be ruled out in this process as there is the process of aromatization is quite a general phenomenon for a variety of substituents." Our structural assignment is further confirmed by a close possibility of vibrationally excited 3 rearranging to 5 and 6 1: 1 correspondence of melting points of the three isomeric benzene devia the sequence 3 ---, 4 --c 7. This di-r-methane-like rearrivatives which can be derived from l and 2: rangement is well documented for h e ~ a m e t h y l p r i s m a n e . ' ~ (H. H ) (H, CH3) If this explanation holds, then generation of 3 via 9 should ortho 193" 186" also yield 5 and 6. 220" 226" meta Dara 270" 269 -270' However, when the experimental test was made, only 6
+
Communications to the Editor
3224 (10) D. Seyferth, C. Sarafidis, and A. B. Evnin, J. Organomet. Chem., 2, 417 (1964). ( 1 1) R. Weiss and S. Andrae, in preparation. (12) Benzene derivatives 5, 6, and 8 are photostable under our irradiation conditions. (13) Professor R. Breslow has kindly informed us that prompted by our results he reinvestigated his earlier work' and was indeed able to confirm his original results. At present the reasons for the discrepancy between his and our results are not fully understood. (14) D.M. Lemai and J. P. Lokensgard. J. Am. Chem. SOC.,88, 5934 (1966). (15) Attempts to observe 3 directly in the photolysis of 9 were unsuccessful. (16) R. Weiss and H. Kdbl, following paper in this issue.
Robert Weiss,* Heinz Kolbl Institut f u r Organische Chemie der Universitat Miinchen 8 Miinchen 2, Germany Received October 2.5. I974
Sir:
So far there have been no reports of Cope rearrangements involving cyclopropene moieties. In this communication we wish to report the first examples, both photochemibicyclopropenyl cal and thermal, of bicyclopropenyl rearrangements which, apart from adding new mechanistic and structural facets to the Cope rearrangement, also allow benzene rearrangesome insight into bicyclopropenyl ment pathways. Direct 320-nm irradiation of la3 (see Scheme I) in 3:1 CH30H-C6H6 (Graentzel reactor 400, 600 w , substrate concentration mol/l., irradiation time 7 hr) gave two new products, 2 (95%) and 3 ( 5 % ) , which after total conversion of la could be separated by fractional crystallization. The structure of 3 (mp 230') follows from spectral comparison with an authentic ample.^ The structural assignment of 2 (mp 173') as a bicyclopropenyl isomeric with l a is based on the following data: ir (KBr) 1845 cm-I ( u
-
-
Scheme I
Ph
\
Ph
y
4
Rhv = CH
Ph
-
2 is photostable under 320-nm irradiation because the 2 is connected with a pronounced blue transformation l a shift in the longest wavelength absorption (331 270 nm). Similarly, under the same irradiation conditions, photolysis of l b gave a mixture of 5 ( 6 5 % ) ,6 (30%), and 7 (5%).2 5 was isolated by low-temperature fractional crystallization from ether and had the following: m p 128-130'; ir (KBr) 269 nm 1845, 1755 cm-' ( u >C=C
